1. Field of the Invention
The present invention relates to an information display system for displaying information related to an object in an image, and a portable information terminal for use in such an information display system.
2. Description of the Related Art
As seen in the broadcasting programs, it has been a common practice to display a computer graphic (CG) image, a telop or the like in superposed relation with an actual video image using a technique such as a chroma key. An image sequence such as an animation is also displayed in superposed fashion. A system for realizing this operation is configured of image input means (camera), means for reproducing a CG or animation, and image superposition means (chroma key or the like). This system makes it possible to display additional information on an actual image.
Also, Jpn. Pat. Appin. KOKAI Publication No. 2001-126051 discloses an apparatus, for example, for displaying the related information in accordance with a view point of the image input means. This apparatus comprises image input means, position/orientation detection means for detecting the position and orientation of the image input means, search and extraction means for searching and extracting information related to a view field of the image input means, and superposed display means for displaying by superposition the input image obtained from the image input means and the related information obtained by the search-extraction means. This configuration makes it possible to display the information in accordance with the view field of the image input means.
Further, a landscape labeling apparatus and system are disclosed in U.S. Pat. No. 6,222,583, for example, in which information related to an image from image input means is displayed. In this system, an approximate position of image input means (image acquisition unit) is acquired from a position information acquisition unit and an angle is acquired from a camera attribute information acquisition unit. A vista image of a view field visible from a particular position is produced in CG based on three-dimensional map information and pattern-matched with the input image thereby to determine the related information superposed.
Also, a method of determining the position and orientation of image input means from a reference mark position is disclosed in “A High Accuracy Realtime 3D Measuring Method of Marker for VR Interface by Monocular Vision”, 3D Image Conference '96, pp. 67–172, by Akira Takahashi, Ikuo Ishii, Hideo Makino, and Makoto Nakashizuka. Specifically, in FIG. 1, assume that Oc is a view point, p0 is an intersection of projected image diagonals and P0 is an intersection of rectangular diagonals, the orientation can be determined by acquiring x′ minimizing
      F    ⁡          (              x        ′            )        =      4    -                  ∑                  i          =          1                4            ⁢                        (                                    n              i                        ⁢                          n              i              ′                                )                2            where the variable vector x′ is given asx′=(α1′β1′,α2′)With regard to a position of image pickup means, on the other hand, a value d minimizing a distance from Ocpi to the apex Pi is estimated by the least-square method.
  d  =                                          (                                          s                0                            ⁢                              s                i                                      )                    ⁢                      (                                          s                i                            ⁢                              w                i                                      )                          -                  (                                    s              0                        ⁢                          w              i                                )                            1        -                              (                                          s                0                            ⁢                              s                i                                      )                    2                      ⁢    D  Then, the position can be determined from an average of a depth value determined from each apex.